Many computing devices, such as desktop computers, workstations, main-frame and super-computers employ multiple printed circuit boards ("PCB") that include various microprocessors, printed circuits and other components that must be electrically coupled together to transmit data and/or power. The electrical traces on one or more layers of the PCB form the printed circuits and typically terminate in one or more terminals or contacts for making connections. Every decreasing element sizes, such a pitch (i.e., the spacing between successive components), width, and height, exacerbate the problem of providing secure and reliable connections between the printed circuits. Precise positioning on the order of thousandths of an inch is often necessary. Consistent pressure across each of the many contacts is also desirable to assure a reliable connection. A single failed or intermittent connection can result in large amounts of "down-time" for the computing device, and costly troubleshooting by highly skilled technicians.
Highly parallel processing super-computers present a particularly significant problem in terms of space constraints. Super computers rely on a high number of connections between circuit boards that each carry one or more microprocessors. The nature of parallel processing places high demands on the timing of signals, including clock signals across the various computer components. The PCBs are spaced relatively close together to reduce the length of the connections between the PCBs in an effort to improve the timing of the signals. The tight spacing hinders the ability of technicians to access particular computer components, such as the PCBs and electrical connectors. This presents a particular problem to computer manufacturers and owners who desire a modular design that permits failed components to be quickly and easily replaced. If serviceable, a modular design would also permit the addition of new or additional processors as desired, for example when more processing power is required or when the processors become more affordable. This could significantly extend the life of the computing device.